Summary: During FY18, we made progress on several subprojects: Our interest in virus assembly is twofold. We seek to identify features with potential for antiviral drug and vaccine interventions, and to elucidate mechanisms at play in the assembly, maturation, and activation of large macromolecular complexes in general. In this context, we work on several viral systems with particular focus on viruses with genomes of double-stranded DNA. These viruses have the largest viral genomes and correspondingly elaborate virion structures. We also have a major interest in hepatitis B virus, a major human pathogen, and in retroviruses, primarily HIV-1. The latter (HIV-related studies) is the subject of a separate report (AR041166-11). During FY18, we made progress on several subprojects: 1) Herpesvirus assembly. Over the past 25 years, we have studied many aspects of herpesvirus assembly, mostly on herpes simplex virus type 1(HSV-1). In recent years, our focus shifted to nuclear egress. The HSV-1 capsid assembles in the nucleus, then migrates into the cytoplasm for subsequent steps on the pathway. The Primary Enveloped Virion (PEV) is a transient particle formed in the perinuclear space as the DNA-filled capsid traverses the nuclear envelope. As the capsid buds through the inner nuclear membrane, it becomes coated with nuclear egress complex (NEC) protein. This yields a PEV whose envelope fuses with the outer nuclear membrane, releasing the capsid into the cytoplasm. To obtain enough PEVs for structural analysis, we isolated them from US3-null-infected cells (pUS3 is a virally encoded kinase). We found that PEVs differ from mature extracellular virions (MEVs) in several respects. PEVs have very few glycoprotein spikes whereas MEVs are densely coated with them. PEVs are 20% smaller than MEVs and there is little space between the capsid and the NEC layer, whereas in an MEV, this space is more capacious and is occupied by the tegument. In FY18, we sought to further characterize the portal vertex. To do so, we worked with A-capsids (empty capsids in a mature conformational state), using a UL25-null strain which yields a high percentage of A-capsids. Tomographic data were recorded and, using a processing algorithm developed in the LSBR, it was possible to identify the portal vertex on these capsids. The portal protein pUL6 is seen to be embedded in the capsid floor of the capsid. In these reconstructions, we also detected a sizeable density overlying the portal vertex that may represent the viral terminase complex. 2) Packing of DNA and internal proteins in bacteriophage T4. In earlier work, our cryo-EM studies of bacteriophage T7 provided strong evidence for the concentric spool model whereby the DNA is coiled around the portal axis in nested shells (Cerritelli et al., Cell 91, 271-290 1997). We have now investigated whether a similar arrangement pertains in T4, which has a 4-fold larger genome, a prolate capsid, and lacks the cylindrical protein core of T7. Cryo-EM yielded side-views and axial views (defined as such relative to the axis running through the portal vertex) of prolate (wild-type) heads and of isometric (mutant) heads. Giant heads were imaged in sideview. We conclude that in isometric heads, the DNA is spooled around the portal axis, essentially as in T7; in giant heads, the spool is rotated by 90o. The innermost regions of DNA-filled heads are less ordered. We infer that the conformations of encapsidated genomes represent an energy-minimized compromise between electrostatic repulsion effects involving DNA duplexes and the strain associated with bending the DNA. Although T4 does not emulate T7 in having a highly ordered core of internal proteins, it does have substantial amounts of internal proteins. In both systems, these proteins are destined for delivery into a host cell whose locations in the head have been unclear. To investigate the internal proteins of T4, of which there are two, gpAlt and gpIPIII, we have used bubblegram imaging, a technique invented in the LSBR (Wu et al., Science 335, 182 2012). Sustained exposure of ice-embedded specimens to the electron beam elicits the formation of bubbles of hydrogen gas in the proteins which are readily detected. In this way we found that both gpAlt and gpIPII are excluded from a highly ordered peripheral zone but otherwise are distributed seemingly at random throughout the capsid interior. The peripheral zone coincides with that occupied by shells of coiled DNA (see above). 3) DNA packaging into supersized T=7 capsids of a thermophilic virus. This project has been a joint undertaking with F. Antson (University of York) as part of a Wellcome Trust/NIH collaborative program. Double-stranded DNA viruses including bacteriophages and herpesviruses package their genomes into preformed capsids, using ATP-driven motors. Seeking to advance structural and mechanistic understanding, we established in vitro packaging for a thermostable bacteriophage, P23-45 of Thermus thermophilus. Both the unexpanded procapsid and the expanded capsid can package DNA in the presence of large terminase and ATP. Cryo- EM reconstructions were determined to 0.4 nm resolution for both the expanded capsid and the procapsid. Unexpectedly, the capsid was found to observe T=7 quasi-symmetry, despite the P23-45 genome being twice as large as those of T=7 and other phages (e.g. HK97, P22) with the same T-number (7) in which DNA packing density is thought to approach the maximum physically possible. Our reconstructions explain this anomaly, whereby modifications to the canonical HK97 fold permit a capsid volume twice as large. We also obtained a 1.95 crystal structure for the portal protein and performed symmetry-free reconstructions for the procapsid and expanded capsid to define its setting in the portal vertex. We found that capsid expansion elicits a substantial change in the conformation of the portal protein, while still allowing DNA to be packaged. 4) Hepatitis B virus is a major cause of acute and chronic liver disease. HBV is a small, enveloped, DNA virus whose core gene codes for two variants: core antigen (cAg), which assembles into capsids; and a precore protein, which is secreted as the non-particulate e-antigen (eAg). The cAg subunit consists of an assembly domain (residues 1-149) and an Arg-rich C-terminal domain (residues 150-183). eAg is directed to the ER by a 29-residue signal peptide. In mature eAg, the assembly domain retains 10 residues of the signal peptide. Previously we showed that this propeptide causes a radically altered mode of dimerization for eAg relative to cAg. In FY17 we carried out structural studies in which a scFv was used as a crystallization chaperone, yielding a co-crystal that diffracted to 0.173 nm. Compared with our earlier structure at 0.32 nm, the new structure shows distinct conformational differences between the paired eAg monomers, indicating a flexibility that may have functional implications. This study was completed and published in FY18. We also investigated the immunological profile of eAg, producing a panel of chimeric rabbit/human monoclonal Fabs selected by phage display. These Fabs were expressed in E. coli, purified, and characterized: some had unprecedentedly high binding affinities and high specificity. The Fabs were used to develop a new and quantitative ELISA-based assay for eAg.